Build agents to learn from experiences using Amazon Bedrock AgentCore episodic memory

Build agents to learn from experiences using Amazon Bedrock AgentCore episodic memory

⋅

Today, most agents operate only on what’s visible in the current interaction: they can access facts and knowledge, but they can’t remember how they solved similar problems before or why certain approaches worked or failed. This creates a significant gap in their ability to learn and improve over time. Amazon Bedrock AgentCore episodic memory addresses this limitation by capturing and surfacing experience-level knowledge for AI agents. Although semantic memory helps an agent remember what it knows, episodic memory […]

Ver mais

Liked Liked
US House takes first step toward creating "commercial" deep space program

US House takes first step toward creating “commercial” deep space program

⋅

A US House committee with oversight of NASA unanimously passed a “reauthorization” act for the space agency on Wednesday. The legislation must still be approved by the full House before being sent to the Senate, which may take up consideration later this month. Congress passes such reauthorization bills every couple of years, providing the space agency with a general sense of the direction legislators want to see NASA go. They are distinct from appropriations bills, which provide actual […]

Ver mais

Liked Liked
Height of harmonic numbers

Height of harmonic numbers

⋅

The previous post looked at writing the harmonic numbers as reduced fractions and estimating the number of digits in the numerator and denominator based on asymptotics. This is a follow up post with plots. We’ll choose our base b to be 2. And we’ll look at the total number of bits in both the numerator and denominator, which we will use as the height of the fractions. First, let’s look at the actual and estimated heights, using the estimates […]

Ver mais

Liked Liked

FastWhisper: Adaptive Self-knowledge Distillation for Real-time Automatic Speech Recognition

⋅

arXiv:2601.19919v1 Announce Type: new Abstract: Knowledge distillation is one of the most effective methods for model compression. Previous studies have focused on the student model effectively training the predictive distribution of the teacher model. However, during training, the student model may inherit the shortcomings of the teacher model, which can lead to a decline in generalization capacity. To mitigate this issue, we propose adaptive self-knowledge distillation (ASKD), which dynamically reduces the dependence of the teacher model to improve […]

Ver mais

Liked Liked

Why You Know What to Do… and Still Don’t Do It

⋅

There’s a question you’ve asked yourself a thousand times. Why is it that you know exactly what you should be doing… and you still don’t do it? You think it’s procrastination. n Or discipline. n Or laziness. The classic explanations we reach for when nothing else makes sense. Here’s the deal, my friend. You don’t have a thinking problem. You have a nervous system problem. Your mind decides what it wants. n Your body decides what you’re allowed to […]

Ver mais

Liked Liked

Guaranteed Nonconvex Low-Rank Tensor Estimation via Scaled Gradient Descent

⋅

arXiv:2501.01696v2 Announce Type: replace Abstract: Tensors, which give a faithful and effective representation to deliver the intrinsic structure of multi-dimensional data, play a crucial role in an increasing number of signal processing and machine learning problems. However, tensor data are often accompanied by arbitrary signal corruptions, including missing entries and sparse noise. A fundamental challenge is to reliably extract the meaningful information from corrupted tensor data in a statistically and computationally efficient manner. This paper develops a scaled […]

Ver mais

Liked Liked

Spectral Diffusion Models on the Sphere

⋅

arXiv:2601.20498v1 Announce Type: cross Abstract: Diffusion models provide a principled framework for generative modeling via stochastic differential equations and time-reversed dynamics. Extending spectral diffusion approaches to spherical data, however, raises nontrivial geometric and stochastic issues that are absent in the Euclidean setting. In this work, we develop a diffusion modeling framework defined directly on finite-dimensional spherical harmonic representations of real-valued functions on the sphere. We show that the spherical discrete Fourier transform maps spatial Brownian motion to a […]

Ver mais

Liked Liked

CrisiSense-RAG: Crisis Sensing Multimodal Retrieval-Augmented Generation for Rapid Disaster Impact Assessment

⋅

arXiv:2602.13239v1 Announce Type: new Abstract: Timely and spatially resolved disaster impact assessment is essential for effective emergency response. However, automated methods typically struggle with temporal asynchrony. Real-time human reports capture peak hazard conditions while high-resolution satellite imagery is frequently acquired after peak conditions. This often reflects flood recession rather than maximum extent. Naive fusion of these misaligned streams can yield dangerous underestimates when post-event imagery overrides documented peak flooding. We present CrisiSense-RAG, which is a multimodal retrieval-augmented generation […]

Ver mais

Liked Liked

Less Data, Faster Convergence: Goal-Driven Data Optimization for Multimodal Instruction Tuning

⋅

arXiv:2603.12478v1 Announce Type: new Abstract: Multimodal instruction tuning is often compute-inefficient because training budgets are spread across large mixed image-video pools whose utility is highly uneven. We present Goal-Driven Data Optimization (GDO), a framework that computes six sample descriptors for each candidate and constructs optimized 1$times$ training subsets for different goals. Under a fixed one-epoch Qwen3-VL-8B-Instruct training and evaluation recipe on 8 H20 GPUs, GDO uses far fewer training samples than the Uni-10x baseline while converging faster and […]

Ver mais

Liked Liked

Load More

↑

Digitado © 2026